Nanoarchitectures for efficient IgE cross-linking on effector cells to study amoxicillin allergy

Allergy. 2021;76:3183–3193. | 3183wileyonlinelibrary.com/journal/all Received: 23 December 2020 | Revised: 28 February 2021 | Accepted: 14 March 2021 DOI: 10.1111/all.14834 ORIGINAL ARTICLE Drug Allergy, Insect Sting Allergy, and Anaphylaxis Nanoarchitectures for efficient IgE crosslinking on effector cells to study amoxicillin allergy Amene Tesfaye1,2 | Alba RodríguezNogales1,2 | Sara Benedé3 | Tahía D. Fernández2,4 | Juan L. Paris1,2 | Maria J. Rodriguez1,2 | Isabel M. JiménezSánchez1,2 | Gador Bogas2,5 | Cristobalina Mayorga1,2,5 | María J. Torres1,2,5,6 | María I. Montañez1,2 This is an open access article under the terms of the Creative Commons AttributionNonCommercialNoDerivs License, which permits use and distribution in any medium, provided the original work is properly cited, the use is noncommercial and no modifications or adaptations are made. © 2021 EAACI and John Wiley and Sons A/S. Published by John Wiley and Sons Ltd. Torres and Montañez contributed equally to this work. Abbreviations: AX, amoxicillin; AXO, amoxicilloyl; AXOBu, amoxicilloylbutylamine; BiAn, bidendron antigens; DNP, dinitrophenyl; HSA, human serum albumin; Hum, human; MCs, mast cells; MoAbs, monoclonal antibodies; MW, molecular weight; PAMAM, polyamidoamine; PEG, polyethylene glycol; PLL, polyLlysine; RAST, radio allergosorbent test; RBL, rat basophilic leukemia cell; SAR, structure– activity relationship; TEM, transmission electron microscopy. 1Andalusian Centre for Nanomedicine and BiotechnologyBIONAND, Málaga, Spain 2Allergy Research Group, Instituto de Investigación Biomédica de MálagaIBIMA, Málaga, Spain 3Instituto de Investigación en Ciencias de la Alimentación (CIAL, CSICUAM), Madrid, Spain 4Departamento de Biología Celular Genética y Fisiología, Facultad de Ciencias, Universidad de Málaga, Málaga, Spain 5Allergy Unit, Hospital Regional Universitario de Málaga, Málaga, Spain 6Departamento de Medicina, Facultad de Medicina, Universidad de Málaga, Málaga, Spain Correspondence María I. Montañez, Andalusian Center for Nanomedicine and BiotechnologyBIONAND, Parque Tecnológico de Andalucía, 29590 Málaga, Spain. Email: maribel.monta[email protected] Funding information European Union’s H2020 research and innovation programme under the Marie SkłodowskaCurie, Grant/Award Number: 713721; Andalusian Regional Ministry Health, Grant/Award Number: RC00042021, PI06992011 and PI01792014; Institute of Health “Carlos III”, Grant/Award Number: PI12/02529, PI15/01206, PI18/00095, CP15/00103, Abstract Background: Amoxicillin (AX) is nowadays the βlactam that more frequently induces immediate allergic reactions. Nevertheless, diagnosis of AX allergy is occasionally challenging due to risky in vivo tests and nonoptimal sensitivity of in vitro tests. AX requires protein haptenation to form multivalent conjugates with increased size to be immunogenic. Knowing adduct structural features for promoting effector cell activation would help to improve in vitro tests. We aimed to identify the optimal structural requirement in specific cellular degranulation to AX using wellprecised nanoarchitectures of different lengths. Method: We constructed eight Bidendron Antigens (BiAns) based on polyethylene glycol (PEG) linkers of different lengths (600– 12,000 Da), endcoupled with polyamidoamine dendrons that were terminally multifunctionalized with amoxicilloyl (AXO). In vitro IgE recognition was studied by competitive radioallergosorbent test (RAST) and antibody– nanoarchitecture complexes by transmission electron microscopy (TEM). Their allergenic activity was evaluated using bone marrowderived mast cells (MCs) passively sensitized with mouse monoclonal IgE against AX and humanized RBL2H3 cells sensitized with polyclonal antibodies from sera of AXallergic patients. Results: All BiAns were recognized by AXsIgE. Dosedependent activation responses were observed in both cellular assays, only with longer structures, containing spacers in the range of PEG 6000– 12,000 Da. Consistently, greater proportion of immunocomplexes and number of antibodies per complex for longer BiAns were visualized by TEM. Conclusions: BiAns are valuable platforms to study the mechanism of effector cell activation. These nanomolecular tools have demonstrated the importance of the adduct size to promote effector cell activation in AX allergy, which will impact for improving in vitro diagnostics. 3184 | TESFAYE ET Al. 1 | INTRODUCTION Drug allergy accounts for 5– 10% of all adverse drug reactions and could result in lifethreatening complications.1,2 βlactam antibiotics are the most frequent triggers of reactions, with amoxicillin (AX) as the most common elicitor nowadays.3 βlactam allergy can be induced by different immune mechanisms, among which the IgE antibodymediated one is the most common and better studied.4,5 The diagnosis of immediate reactions to AX is based mainly on in vivo and in vitro methods,6 being drug provocation test the gold standard, although it is risky and not recommended in patients with severe reactions.2,7,8 In vitro tests are based on the determination of specific IgE (sIgE), with the commercial ImmunoCAP only detecting 20% of allergic patients, and on the quantification of basophil activation after stimulation with the culprit drug,9,10 showing sensitivity around 50%. Among the factors affecting such low sensitivity could be the fact that correct antigenic determinant and/or conjugates are not incorporated to in vitro tests.6,11,12 According to the hapten hypothesis, AX is a low molecular weight (MW) compound that must form protein covalent conjugates with increased size and multivalence to be immunogenic.13 Spontaneous conjugation of AX occurs due to the structural propensity of its βlactam ring opening by the nucleophilic primary amines from proteins that results in the amoxicilloyl (AXO) antigenic determinant.14 The immunological recognition of such multivalently presented antigenic determinants on a conjugate by, at least, two adjacent IgE antibodies that are bound to their highaffinity receptor (FcεRI) on the surface of tissue mast cells (MCs) or circulating basophils, results in an intricate process of IgE crosslinking,15– 18 releasing preformed inflammatory mediators and eliciting the acute allergic response.15,19 The efficiency of the stimulation on cell degranulation is dependent on many factors, including the drug antigenic determinant structure,18,20 its valency on the conjugate or complete antigen,21,22 the size of the conjugate,15,23 the proximity of the IgE epitopes,24 and the steric hindrance.1 6 , 2 4 – 2 6 The study of these complex cellular and structural restrictions in the in vitro activation of effector cells requires sophisticated structures that are welldefined and characterized to facilitate the interpretation of results. In this regard, different molecules have been designed to assess the influence of these parameters on the degranulation of MCs using synthetic haptens linked on bior multivalent structures. The most widely used experimental model utilizes the rat basophilic leukemia cell (RBL)/dinitrophenyl (DNP) and/or dansyl system, in which the RBL cells are primed with monoclonal IgEDNP and/or IgEdansyl antibodies, followed by stimulation with a multivalent haptenated structure to induce degranulation.15,16,18,23,26– 30 Dissimilar optimal sizes have been reported for effective crosslinking as a consequence of the diverse features of the synthetic allergen systems studied, differing in the density of haptens (or valency),23,30 threedimensional structure, flexibility23, or rigidity15,29,30 of the structures. KEYWORDS amoxicillin, drug allergy, IgE crosslinking, immunocomplex, nanostructure GRAPHICAL ABSTRACT The use of BiAns demonstrates the importance of adduct size and distance between determinants to promote effector cell activation in AX allergy. Optimal effector cell activation is showed with the biggest BiAns, which involves a greater number of immunocomplex and antibodies. BiAns are versatile nanoplatforms that can be applied to different allergies, valuable for improving in vitro allergy tests. Abbreviations: AX, amoxicillin; AXO, amoxicilloyl; BiAn, bidendron antigen; MoAb, monoclonal antibody; RBL, rat basophilic leukemia cell PI17/01237, PI20/01734, PI20/01447, ARADYAL RD16/0006/0001 and Euronanomed Program AC19/00082 | 3185 TESFAYE ET Al. However, owing to the complexity of real drug allergy scenario, this kind of structure– activity relationship (SAR) approach with nanostructures has never been addressed involving drugs (as haptens) or samples from allergic patients (drugsIgE). Inspired by our previous design of dendrimeric antigens that permit controlling the size, multivalence (number of haptens), and threedimensional structure,31,32 and showed potential use for drugsIgE quantification,33– 37 we propose a new related nanoarchitecture design to expand the interaction studies with the immune system. Herein, we construct a set of symmetrical dendrimerderived nanoarchitectures, called Bidendron Antigens (BiAns), in which two dendrons decorated with multiple units of AXO are separated by flexible polymeric spacers of different lengths (Figure 1), to evaluate the influence of the nanoarchitecture size for promoting the activation of effector cells that causes drug allergic reactions. This SAR study will help understand deeply the required distance between AXO determinants to activate basophils or MCs, which would be useful for improving the sensitivity of in vitro tests for diagnosing allergic reactions to AX. FIGURE 1 General synthetic scheme including the following: (A) the intermediate PEG diamine; (B) the intermediate PEGbismaleimide; and (C) the final BiAns, nanoarchitectures bearing 16 AXO determinants. The structural variation among different BiAns depends on the PEG length (with n ranging from 14 to 273). Briefly, the synthetic procedure consisted in obtaining bismaleimideactivated PEG compounds that allowed coupling between thiolcore dendrons (G2), whose peripheral groups were eventually functionalized with AX. In building up this, series of independent chemical reactions were followed, and the completion of the reactions was monitored by 1H and 13C NMR through the appearances and disappearances of distinctive signals. Reaction conditions: (a) TsCl, CH2Cl2, KOH, 0°C; (b) NaN3, THFH2O (80:20), reflux at 70°C; (c) H2, Pd/C, MeOH, (d) iBuCOCl, NMM, CH2Cl2, 0°C; (e) PEG- (NH2)2, CH2Cl2; (f) PBS, pH 6.5, TCEP; (g) PEGbismaleimide, H2O, DMSO; (h) AX, carbonate buffer (0.05M, pH 10) [Color figure can be viewed at wileyonlinelibrary.com] 3186 | TESFAYE ET Al. 2 | METHODS 2.1 | Patients Patients with a clinical history of immediate allergy to AX and tolerant subjects were evaluated following the European Academy of Allergy and Clinical Immunology (EAACI) guidelines.38,39 The study was conducted according to the Declaration of Helsinki principles and was approved by the Provincial Ethics Committee of Malaga. All subjects included in the study were informed orally and signed the corresponding informed consent. For evaluation of AXsIgE recognition, we used a pool of 14 sera from patients with positive skin tests to AX and high levels (>7%) of AXsIgE measured by direct radioallergosorbent test (RAST). Patients’ data are displayed in Table S1. Sera with high levels of total IgE, measured by ImmunoCAP (≥250 kU/L) were selected for humanized RBL2H3 (HumRBL2H3) cell evaluation studies: three patients allergic to AX, with high levels (>7%) of AXsIgE measured by RAST (data shown in Table S2), and three tolerant subjects, with no reactivity to βl a c t a m s . 2.2 | Production of Conjugates and AntiAXO monoclonal antibodies (MoAbs) The synthesis, purification, and characterization of the conjugate of human serum albumin (HSA) with AX (HSAAXO) and the series of BiAn and intermediates, and the production of two hybridomae:40 AO3.2, IgG2a isotype, and AO6.2, IgE isotype are detailed in the Methods section of this article’s Online Repository. 2.3 | Radioimmunoassays Radioimmunoassays were performed by RAST as previously described14 using the polyllysine (PLL) cellulose solidphase conjugated to AX (AXOPLL cellulose disks) and 125I - a n t i - I g E . 41 Results were considered positive if they were higher than 2.5% of label uptake, which was the mean ± 2 SD of the negative control group. Competitive inhibition immunoassays were carried out with a pool of sera with high RAST values (>7%), which was incubated with the inhibitors (AXObutylamine (AXOBu) and BiAns) at four concentrations (30, 15, 3, and 0.3 mM of AXO equivalents) as the fluid phase for 18h at room temperature. Then, AXOPLL solidphase disks were added, followed by the described RAST protocol. Results were calculated as the percentage of inhibition using the noninhibited serum as a control. 2.4 | Transmission electron microscopy The samples were prepared according to optimized negative staining (OpNS).42 From the different tested MoAbs to BiAn ratios, a molar ratio of 8:1 was found to be the optimal condition. This ratio was used to incubate MoAbs with BiAn(600), BiAn(6000), and BiAn(10000) to evaluate differences in the obtained complexes as a function of the employed nanoarchitecture. 2.5 | Cell viability and activation assays Viability was assessed using 3- (4,5dimethylthiazol2yl)- 5- (3carb oxymethoxyphenyl)- 2- (4sulfophenyl)- 2Htetrazolium (MTS) assay. Bone marrow cells were collected from femurs of BALB/c mice (4– 6 weeks of age) from Charles River Laboratories (Saint Germain sur l´Arbresle) and cultured in DMEM with glucose and lglutamine, fetal bovine serum, penicillin/streptomycin, and sodium pyruvate (all from Gibco, Life Technologies), plus recombinant murine SCF, IL3, IL9, and TGFβ (all cytokines and growth factors were from Peprotech, Rocky Hill, NJ) to differentiate into MCs as previously described.43 MCs were cultured for a minimum of 4 weeks and up to 8 weeks before they were used for functional assays. For activation through crosslinking of the IgE receptor, MCs were initially sensitized for 4h with 1 µg/mL of mouse antiAX IgE MoAb. After washing, MCs were resuspended in Tyrode´s buffer and activated with the BiAns at 10, 50, and 100 µM (concentration of AXO units) for 1h. A conjugate of HSA with AX at 50 μg/ml (0.7μM of HSAAXO, corresponding to 9, 7 μM of AXO) was used as a positive control. Control experiments were performed in unsensitized cells. Experiments were performed in triplicate. All protocols involving animals followed the European legislation (Directive 2010/63/EU) and were approved by Comunidad de Madrid (Ref PROEX 089/15). HumRBL2H3 cells were obtained from the Cell Culture Unit of the University of Granada (Granada, Spain) and cultured in RPMI containing 10% heatinactivated FBS, 2mM LGlutamine, 100 IU/ml penicillin, and 100 µg/ml streptomycin, in a humidified 5% CO2 atmosphere at 37°C. HumRBL2H3 cells were seeded in 96well plates at a density of 2 × 105 cells/well. Confluent growing HumRBL2H3 cells were then sensitized with serum (50% v/v) from AXallergic patients (n = 3) and tolerant subjects (n = 3) for 48 h at 37°C. Unsensitized cells were used as controls. Subsequently, cells were stimulated with the series of BiAns at different concentrations of AXO units (1, 10, 20, 50, 100, 150, and 200 μM) for 2 h. In parallel experiments, cells were stimulated with the corresponding PEG linkers (structures without dendronAXO), as negative controls. The HSAAXO conjugate (at 10 μM concentration of AXO) was used as a positive control of degranulation induction. 2.6 | Quantification of βhexosaminidase Release The cell degranulation response was quantified measuring the level of βhexosaminidase released in culture supernatants.44 The percentage of βhexosaminidase release was calculated as a percentage of the total βhexosaminidase content. For further details, see the Methods section of this article’s Online Repository. | 3187 TESFAYE ET Al. 3 | RESULTS 3.1 | Design of Dendrimerbased Nanoarchitectures A novel set of symmetrical dendrimerderived nanoarchitectures, BiAns, was designed, synthesized (Figure 1), purified, and characterized (Figure S1). The BiAns were constructed to display critical design features: (a) bearing multiple units of AX antigenic determinant on the periphery of the dendron scaffolds to support multivalent interaction with the target antibody; (b) a flexible hydrophilic PEG spacer that provides a specific length between dendrons, promotes aqueous solubility,45 and reduces nonspecific protein adsorption, immunogenicity, and antigenicity.23,46 A series of PEG chains of eight different average MWs (range 600– 12,000 Da) were chosen as linkers. Commercially available secondgeneration (G2) polyamidoamine (PAMAM) dendron was selected as a common multivalent scaffold as it displays peripheral amino groups that covalently bind to AX in an efficient manner, and previous studies showed that dendrimeric antigens based on PAMAM are recognized by penicillinsIgE.32,33,35,36 The resulting BiAns encompass the same number of peripheral functionality (16 AXO per BiAn), differing only in the length of the hydrophilic spacer separating the dendrons, ranging in aqueous solution from 3 to 20 nm, with a maximal possible extension from 4.8 to 95.5 nm (Table 1). 3.2 | IgE Recognition of BiAns The ability of AXsIgE to recognize BiAn conjugates was evaluated by competitive RAST inhibition immunoassays using sera from AXallergic patients. This assay consists in competitive serum IgE recognition between the solid phase (AXOPLL conjugate attached to cellulose disks) and the different inhibitors, containing AXO at different concentrations, in the fluid phase, and results are represented as % inhibition IgE binding (Figure 2). A butylamineAXO monomeric conjugate was also employed as a control inhibitor.14 Assays were performed at equimolar amounts of AXO for all the conjugates, with a maximum concentration not higher than 30 mM of AXO, due to solubility issues at higher concentrations of BiAns. At such maximum concentration, sera were inhibited by all BiAns. In general, inhibition dropped at 3 and 0.3 mM of AXO in inhibitors. The presented data support that all compounds are recognized by AXsIgE in a concentrationdependent manner, with no differences on the degree of IgE recognition among the different BiAns, as they present the same number of AXO equivalents (Figure 2). 3.3 | Antibody– nanoarchitecture Complexes To visualize IgE binding to the BiAn nanoarchitectures and the shape of the resulting immunocomplex, OpNS TEM was performed.47,48 Optimization of the staining and imaging conditions with MoAb alone, AO3.2 (specific to AXO structure), showed Yshaped monomeric antibodies in different orientations (Figure 3A) and their corresponding two Fab and one Fc domains (Figure 3B). To study the influence of spacer length of the BiAns on the formation of immunocomplexes, various nanoarchitectures (BiAn(600), BiAn(6000), and BiAn(10000)) were incubated with the antibody prior to evaluation of the complexes by OpNS TEM. In all cases, the immunocomplex formation followed a similar pattern, with most of the observed particles corresponding to individual antibodies (60– 83%). Since the attached nanoarchitectures were not detectable, it was not possible to distinguish monomeric complexes from noncomplexed antibodies. Considering only structures with more than one antibody per particle, the most predominant configuration corresponded to dimeric complexes, which appeared to be the most energetically favored structure (Figure 3C, Table S3). Since the length of the PEG spacer in the BiAn increased (PEG 10,000 > 6000 > 600), the percentage of complexes with 2 or more than 2 antibodies also increased. This might be ascribed to a larger steric hindrance in the structures with shorter PEG spacer, where the two AXO dendrons are much TABLE 1 Estimated and experimental values of synthetic BiAns Synthetic antigen Estimated MW (Da) Linker PEG MW (Da) (C2H4O)N where N is Flory radius (nm) Solution PEG length (nm) Extended PEG length: Contour length(nm) D (m2/s) Valency BiAn 600 10,471.9 600 14 1.68 3.36 4.8 1.53 × 10−10 16 BiAn 1000 10,871.9 1000 23 2.28 4.56 8.0 1.32 × 10−10 16 BiAn 2000 11,871.9 2000 45 3.46 6.92 15.9 1.29 × 10−10 16 BiAn 4000 13,871.9 4000 91 5.24 10.48 31.8 1.24 × 10−10 16 BiAn 6000 15,871.9 6000 136 6.68 13.36 47.7 1.09 × 10−10 16 BiAn 8000 17,871.9 8000 182 7.94 15.88 63.6 9.64 × 10−11 16 BiAn 10000 19,871.9 10,000 227 9.08 18.16 79.5 9.57 × 10−11 16 BiAn 12000 21,871.9 12,000 273 10.13 20.26 95.5 9.50 × 10−11 16 Note: The Flory radius (RF) is calculated with RF = a.N3/5 (where a is the length of a monomer unit, and N the number of repeating monomeric units. Solution PEG length is calculated as the diameter of the polymer in aqueous solution based on Flory radius. The contour length (full extended length) is calculated as a product of the polymeric length (N) and length of the monomeric unit, 3.5 Å for PEG.54 3188 | TESFAYE ET Al. closer to each other, preventing the binding of other antibodies to form the immunocomplex. Besides the number of antibodies per complex, other differences could be also detected regarding the morphology of the complexes as a function of the studied molecule (Figure 3D). For complexes involving four antibodies with an open structure, up to three antibodies could be found to bind on a common spot (which could contain both dendrons of one BiAn). For larger complexes, conformational elucidation becomes extremely challenging due to the superposition of several antibodies. It is important to bear in mind that the antibody saturation conditions of the experiment may favor the formation of ringclosed complexes, and therefore minimize crosslinking potential. In this context, it is worth noting that all the complexes observed with BiAn(600) presented an open or linear structure (Figure 3D), which in the case of 2:2 complexes might correlate with the tooshort spacer between both antigenic dendrons that impedes the simultaneous interaction with both recognition sites of a single antibody (Table 1), estimated in 11– 13 nm.49 On the other hand, and although open structures were still predominant, Figure 3D also shows some examples of ringclosed complexes with different numbers of antibodies per complex that were observed with BiAn(6000) (6.3% of the complexes) and BiAn(10000) (6.2%). 3.4 | Toxicity assay Effect of different concentrations of BiAns on HumRBL2H3 cell viability revealed a survival rate greater than 70% at most concentrations assayed. Only the highest one (200 μM) in BiAn(2000) and BiAn(4000) reduced the HumRBL2H3 viability up to 64% (data not shown). 3.5 | Effects of BiAns on IgE activation of bone marrowderived cells The capacity of nanoarchitectures to induce IgEdependent degranulation of bone marrowderived MCs was evaluated. For this, cells were sensitized with IgE MoAb against AX (specific to AX side chain), then treated with different concentrations of BiAns (in terms of the same equivalents of AXO), and afterward, βhexosaminidase assay was performed (Figure 4, left). Stimulation of cells with HSAAXO (10 μM of AXO moieties) induced up to 25% of βhexosaminidase release. However, cell stimulation with BiAns using an equivalent concentration of AXO only caused high activation with the BiAn(10000), inducing 17% of βhexosaminidase release. Only BiAns constructed with PEG of MW ≥6000 Da induced cell degranulation, and in a dosedependent manner, showing up to 19% of βhexosaminidase released at higher concentrations (50 and 100 μM of AXO moieties), and bringing out the importance of the polymeric spacer length to achieve cell activation (Figure 4B). None of the BiAns tested induced cell degranulation on unsensitized cells (Figure 4A), indicating that activation occurs through an IgE mechanism. 3.6 | Effect of BiAns on IgEinduced degranulation in RBL2H3 cells Next, we chose the HumRBL2H3 cell line for evaluation, which shares some characteristics with both MCs and basophils, and expresses human IgE receptor (FcεRI).27 Cells were primed with polyclonal antibodies from sera of AXallergic patients and tolerant subjects, then sensitized RBL2H3 cells were treated with different concentrations of BiAns (in terms of the same equivalents of AXO determinants), and subsequently, βhexosaminidase assay was performed. Results, shown in Figure 4 right, indicate that sensitized cells with sera from allergic patients and stimulated with BiAns containing PEG of MW ≥6000 Da significantly induced the βhexosaminidase release in a concentrationdependent manner, compared with negative control groups (PBS and HSA activated cells) (Figure 4B bottom, Figure S3). Blank structure controls (PEGs) did not induce cell degranulation (Figure S3). Moreover, the highest concentrations of BiAns containing MW ≥6000 Da induced a release of 33% of βhexosaminidase, similar to the one induced by HSAAXO (at 10 µM conc of AXO) (Figure 4B). Among these, BiAn(10000) and BiAn(12000) are the most effective intermolecular FIGURE 2 RAST inhibition assays performed with a pool of sera from patients allergic to AX, using the series of BiAns and a AXOBu conjugate as inhibitors and cellulose disks modified with AXOPLL as the solid phase. Specific IgE recognition is considered with inhibition of ≥50% [Color figure can be viewed at wileyonlinelibrary.com] | 3189 TESFAYE ET Al. crosslinker as they induced above 27% of βhexosaminidase release at 10, 50, and 100 µM of AXO, although at lower concentration (1 µM) only BiAn(10000) induced such substantial degranulation (Figure S3). No significant differences were observed in these BiAns treatments when the cells were sensitized with sera from tolerant subjects (Figure 4B top), and on unsensitized cells included as control of IgE activation (data not shown), whereas a more specific dose– response effect was observed when the cells were sensitized with sera from patients. The fact that none of the BiAns induced cell degranulation on unsensitized cells excludes direct activation by offtarget occupancy of cell surface receptors. In addition, the absence of activation observed in parallel experiments with HMC 1.2 cell line, which exhibited a similar phenotype to that of human MCs, expressing IgG receptor (FcγR) but not the highaffinity IgE receptor (FcεR),50 demonstrated that the BiAns do not trigger degranulation by an IgGmediated pathway (Figure S4) suggesting that BiAns stimulate the degranulation on HumRBL2H3 through an IgE pathway. 4 | DISCUSSION Our SAR study with wellcharacterized BiAns shows that the size and the proximity of AXO determinants on these conjugates influence the number and shape of immunocomplexes and their FIGURE 3 Transmission electron micrograph of negatively stained MoAbs: (A) unbound MoAbs; (B) Zoomedin views of selected individual antibody images. Scale bars represent 10 nm; (C) Number of MoAbs per complex after incubation with BiAn(600), BiAn(6000), or BiAn(10000); (D) Zoomedin views of selected individual complex images showing different open and ringclosed structures. Bars represent 10 nm [Color figure can be viewed at wileyonlinelibrary.com] 3190 | TESFAYE ET Al. subsequent ability to activate in vitro effector cells in an IgEdependent mechanism. Through competitive immunoassays, it could be shown that all BiAns were recognized by AXsIgE from patients. The higher concentration of determinants (AXO) increased the extent of inhibition, a pattern already reported with the monomeric AXOBu conjugate14 and dendrimeric antigens with different penicillin determinants.32 The increased inhibition obtained with the monomeric AXOBu, compared with BiAns, can be attributed to the absence of steric interactions for sIgE binding to AXO moieties in this small conjugate. In BiAn, PEG chains could contribute to a steric hindrance to the IgE binding, and also the high proximity of the eight AXO determinants exposed in each dendron would hinder their simultaneous IgE recognition. Using OpNS TEM, immunocomplexes could be visualized for different BiAn sizes. Assessment of BiAn immunogenicity at cellular level reveals that IgEmediated degranulation of bone marrowderived MCs and HumRBL2H3 cells with BiAns is polymeric spacer length dependent. In both cellular assays, dosedependent activation responses were observed with all BiAn containing flexible linkers above a critical size (PEG 6000). Although these PEG polymers take on a spherical equilibrium configuration in an aqueous environment (for instance 13 nm for PEG 6000 in its folded conformation), their chain units move freely51 and both ends could be at any position within the contour length of the polymer chain (∼48 nm) (Table 1). Similar levels of activation were found between BiAn(10000) or BiAn(12000) compared with HSAAXO activation at the same concentration of AXO in the RBL assays. However, in the case of MC, the same level of activation is only achieved with higher concentrations of BiAns. That could be explained by the better accessibility of AXO moieties on HSAAXO which could allow the crosslinking of more than two IgE with a single molecule. On the other hand, given the proximity among the AXO units attached to the same dendron moiety, it is unlikely that more than two IgE are bound to each BiAn during crosslinking. OpNS TEM data support cellular activation results, since the BiAn that leads to a greater proportion of immunocomplexes, and the largest FIGURE 4 Degranulation assays after incubation of cells with the series of BiAns at 10, 50, and 100 μM of AXO equivalents. HSA and HSAAXO (at 10 μM of AXO) were used as negative and positive control, respectively. (A) Percentage of βhexosaminidase release in unsensitized (top) and sensitized (bottom) bone marrowderived MCs; (B) Percentage of βhexosaminidase released by sensitized cells with sera from tolerant subjects (top) or with sera from AXallergic patients (bottom). Data are expressed as means ± SD. The baseline of percentage of βhexosaminidase release is represented by the dotted horizontal line [Color figure can be viewed at wileyonlinelibrary.com] | 3191 TESFAYE ET Al. number of antibodies per complex, was also the most successful one in activating cellular responses, in agreement with reported potent degranulation achieved by synthetic allergens with a valency ≥3.29,30,52 The failure of BiAn containing PEG ≤4000 to stimulate degranulation indicates that, despite their recognition by AXsIgE, they are inefficient at crosslinking cell surface IgE, which could be explained by a relatively low abundance of extended conformation of the PEG polymers in aqueous solution. BiAn containing the larger PEG length, and bearing equal valency, facilitates this interaction, presenting BiAn(6000) the minimal distance between haptens that can induce cell degranulation. The most potent stimulator BiAn(10000) is effective at all studied concentrations in both cell lines, MC and RBL2H3 (as low as 1 µM for RBL2H3) (Figure S3), indicating that an estimated distance of ∼18 nm between haptens seems to be optimal for crosslinking the receptors. Consistent with the literature, this ∼20 nm dimension was found to be the optimal distance between DNP haptens on a rigid nanoparticle system to induce MC degranulation.30 However, this is not in agreement with the pattern observed in DNP divalent and trivalent systems, in which rigid spacers of 4– 5 nm stimulate stronger degranulation responses compared with those possessing spacing greater than 7– 10 nm.15,29 Our findings indicate that not only the size and multivalence of nanostructures are important factors for inducing degranulation, but also their flexibility. Overall, these results suggest that BiAn nanoarchitectures containing longer PEG chains (MW range: 6000– 12,000 Da) are effective triggers, whereas bivalent structures of DNP, in which haptens are also separated by flexible PEG of different lengths (MW range: 400– 10,000 Da), were reported not to activate MC, but to behave as inhibitors.23 This inhibition was explained by a preferential formation of intramolecular crosslinking of antibody by bivalent DNP of sufficient PEG length (10 nm) (stable 1:1 complexes),23 or formation of cyclic dimmers of IgEFcεRI on the cell surface with shorter linkers (<5 nm).28 Comparisons in terms of chemical structure between BiAn and DNPPEG systems23 points to multivalent vs. bivalent hapten presentation as the main difference. Likely, the multivalence of the dendron in BiAn favored IgE interaction, upon dendritic or synergetic effect, and therefore the degranulation. This is in agreement with cell activation induced by other multivalent systems: dendrimers presenting 16 units of DNPinduced MCs degranulation in DNP studies interaction;23 different penicillin dendrimeric antigens activated basophils from patients, with increased stimulation index observed for those displaying higher valence (64 vs 16 haptens);53 other rigid systems, nanoparticles (≥19.8 nm) functionalized with multiple DNP, showed to be very effective effectors, however, a reduced hapten density inhibited degranulation.30 The scenario for MC and basophil degranulation is very complex, and the use of defined nanoarchitectures has allowed the identification of the minimal requirements for their activation in a realistic model. In this regard, multivalent dendritic presentation and distance between the haptens in BiAn constructed with PEG MW range 6000– 12,000 fulfilled the optimal requirement to overcome the intricate cellular preconditions leading to cell activation. The optimal distance between AXO determinants for effective crosslinking is observed in BiAn(10000). In summary, using multivalent AXO dendrons spaced by flexible PEG polymers, this study sheds light on the mechanism of the effector cells activation from a unique realistic perspective, using human samples and haptens in clinical use. Moreover, the synthesis of BiAn platform is versatile and could apply to different drug haptens or allergen epitopes. Understanding the biology and nanoscale organization of the cell membrane receptors can lead to the development of novel diagnostic and therapeutic tools for drug allergy. ACKNOWLEDGEMENTS The characterization of nanostructures by mono and bidimensional NMR techniques as well as the optimization of TEM studies for visualising immunocomplex has been performed by the ICTS “NANBIOSIS”, more specifically by the U28 Unit of the Andalusian Centre for Nanomedicine & Biotechnology (BIONAND). The present study was supported by the Institute of Health “Carlos III” (ISCIII) of MINECO (grants cofunded by ERDF: “Una manera de hacer Europa” (grant numbers PI12/02529, PI15/01206, CP15/00103, PI17/01237, PI18/00095, PI20/01734, PI20/01447, RETICS ARADYAL RD16/0006/0001, Euronanomed Program AC19/00082); Andalusian Regional Ministry of Economy and Knowledge (grants cofunded by ERDF: “Andalucía se mueve con Europa”: grant no. CTS06603); Andalusian Regional Ministry of Health (grant nos PI06992011, PI01792014); and “Premio UNICAJA a la innovación en biomedicina y salud.” AT has received funding from the European Union’s H2020 research and innovation program under the Marie SkłodowskaCurie (grant no. 713721). S.B. acknowledges financial support of Ministerio de Ciencia, Innovación y Universidades from Spain through a Juan de la Cierva Incorporación contract. C.M. holds “Nicolas Monardes” research contract by Andalusian Regional Ministry Health (grant no. RC00042021). G.B. holds a “Juan Rodes” grant (JR18/00054), J.L.P holds a “Sara Borrell” grant (CD19/00250), A.R.N. and M.I.M. hold a “Miguel Servet I” grant (CP19/00191 and CP15/00103), both grants cofunded by European Social Fund (“El FSE invierte en futuro”). Tesfaye reports grants from Marie SkłodowskaCurie [grant No 713721], during the conduct of the study; in addition, Tesfaye has a patent PCT/ES2021/070103 pending. Dr. RODRIGUEZ NOGALES reports grants from ISCIII, during the conduct of the study; in addition, Dr. RODRIGUEZ NOGALES has a patent PCT/ES2021/070103 pending. Dr. Benede reports grants from Ministerio de Ciencia, Innovación y Universidades, during the conduct of the study. Dr. Fernandez reports grants from ISCIII, during the conduct of the study; In addition, Dr. Fernandez has a patent PCT/ES2021/070103 pending. Dr. Paris reports a Sara Borrell fellowship from ISCIII (CD19/00250), cofunded by European Social Fund. Dr. Rodriguez reports grants from ISCIII, during the conduct of the study. JIMÉNEZSÁNCHEZ reports grants from MICIN (PEJ2018002865A), during the conduct of the study; Dr. BOGAS HERRERA reports grants from ISCIII, during the conduct of the study. Dr. Mayorga reports grants